WO2020134943A1 - 一种车险自动赔付方法和系统 - Google Patents

一种车险自动赔付方法和系统 Download PDF

Info

Publication number
WO2020134943A1
WO2020134943A1 PCT/CN2019/123328 CN2019123328W WO2020134943A1 WO 2020134943 A1 WO2020134943 A1 WO 2020134943A1 CN 2019123328 W CN2019123328 W CN 2019123328W WO 2020134943 A1 WO2020134943 A1 WO 2020134943A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
image
credit
vehicle
insurance
Prior art date
Application number
PCT/CN2019/123328
Other languages
English (en)
French (fr)
Inventor
周凡
Original Assignee
阿里巴巴集团控股有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 阿里巴巴集团控股有限公司 filed Critical 阿里巴巴集团控股有限公司
Publication of WO2020134943A1 publication Critical patent/WO2020134943A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/20—Administration of product repair or maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements

Definitions

  • the present disclosure relates to financial technology, and in particular to a method and system for auto insurance auto payment.
  • the owner will usually call the insurance company to report the incident.
  • the insurance company will arrange for the survey personnel to inspect the damage at the scene, or the owner will take a picture of the damage and send the photo to the insurance company. After the remote survey, determine the items that need to be repaired and replaced, and then the owner will go to the repair shop for repair. After getting the invoice or repair list, these materials will be handed over to the insurance company. After the insurance company has reviewed all relevant documents, it will pay for the repair costs. To the owner. The whole process is relatively lengthy, the owner needs to pay more attention to the matter, and the speed of getting compensation is also very slow.
  • the existing compensation process wants to be able to achieve strict control of the process and reduce the unreasonable compensation ratio, it will make the compensation process longer, cumbersome procedures, and poor owner experience; if you want to reduce the owner's trouble, simplify the process, quickly pay, and will Increase the proportion of unreasonable compensation.
  • this application provides an auto insurance automatic payment solution to solve the above-mentioned defects.
  • this application hopes to provide a method of automatic compensation for pure-looking cases through image recognition, artificial intelligence, deep learning, Internet payment technology and user credit system, which can realize the real-time payment of compensation payments, avoiding the owner and investigation There are too many processes or procedures involved by staff, and they can control the accuracy of compensation to prevent fraud.
  • a method for automatic payment of auto insurance includes:
  • the user credit is reviewed when the user accepts the maintenance list, and the method further includes reacquiring the vehicle image if the user does not accept the maintenance list.
  • the automatic compensation is executed when the user's credit is sufficiently high and the method further includes converting the user's current settlement process to manual processing if the user's credit is not sufficiently high.
  • a system for automatic payment of auto insurance includes:
  • a device for reviewing user credit includes
  • the user credit is reviewed when the user accepts the maintenance list, and the system further includes means for reacquiring the vehicle image if the user does not accept the maintenance list.
  • automatic compensation is executed when the user's credit is sufficiently high, and the system further includes a method for converting the user's current claims process to manual processing when the user's credit is not high enough Device.
  • a computer-readable storage medium that stores instructions for auto insurance auto payment, and the instructions include:
  • the user credit is reviewed when the user accepts the maintenance list, and the computer-readable storage medium also includes instructions for reacquiring the vehicle image if the user does not accept the maintenance list .
  • automatic compensation is executed when the user's credit is sufficiently high, and the computer-readable storage medium further includes a method for transferring the user's current settlement process if the user's credit is not high enough Instructions for manual processing.
  • FIG. 1A, 1B, and 2 illustrate various operating environments in which embodiments of the present application can be implemented.
  • FIG. 3 shows a block diagram of an example of an automobile insurance claim application according to an embodiment of the present application.
  • FIG. 4 shows an example of a graphical user interface presented to a user by a UI presentation module according to an embodiment of the present application.
  • 5A and 5B show another example of an application for auto insurance claims according to another embodiment of the present application.
  • FIG. 6 shows a detailed functional block diagram of an image recognition module according to an embodiment of the present application.
  • FIG. 7A shows an example of reflection removal in image recognition according to an embodiment of the present application.
  • FIG. 7B shows an example of angle correction in image recognition according to an embodiment of the present application.
  • FIG. 8 shows a flowchart of a method for auto insurance auto payment according to an embodiment of the present application.
  • FIG. 9 shows a flowchart of a method for image recognition according to an embodiment of the present application.
  • FIGS. 1A, 1B, 2 and the associated description provide a discussion of various operating environments in which embodiments of the present application can be implemented.
  • the devices and systems shown and discussed with respect to FIGS. 1A, 1B, and 2 are for purposes of illustration and description, rather than extensive calculations that can be used to implement various embodiments of the application described herein Device configuration limitations.
  • FIG. 1A and 1B illustrate suitable mobile computing environments that can be used to implement various embodiments of the present application, such as mobile phones, smart phones, tablet personal computers, laptop computers, and so on.
  • FIG. 1A there is shown an example mobile computing device 100 for implementing various embodiments.
  • the mobile computing device 100 is a handheld computer with both input and output elements.
  • the input elements may include a touch screen display 105 and input buttons 110 that allow a user to input information into the mobile computing device 100.
  • the mobile computing device 100 may also incorporate an optional side input element 115 that allows further user input.
  • the optional side input element 115 may be a rotary switch, a button, or any other type of manual input element.
  • the mobile computing device 100 may incorporate more or fewer input elements.
  • the display 105 may not be a touch screen.
  • the mobile computing device is a portable telephone system, such as a cellular phone with a display 105 and input buttons 110.
  • the mobile computing device 100 may also include an optional keypad 135.
  • the optional keypad 135 may be a physical keypad or a "soft" keypad generated on a touch screen display.
  • the mobile computing device 100 incorporates output elements, such as a display 105 that can display a graphical user interface (GUI). Other output elements include speaker 125 and LED 120.
  • the mobile computing device 100 may include a vibration module (not shown) that causes the mobile computing device 100 to vibrate to notify the user of the event.
  • the mobile computing device 100 may incorporate a headphone jack (not shown) for providing another means to provide an output signal.
  • the present application can also be used in combination with any number of computer systems, such as in a desktop environment, laptop or notebook computer systems, multi-processor systems, based on Microprocessors or programmable consumer electronics, network PCs, small computers, mainframe computers, etc.
  • the embodiments of the present application can also be practiced in a distributed computing environment, where tasks are performed by remote processing devices that are linked through a communication network in the distributed computing environment; programs can be located in local and remote memory storage devices.
  • any computer system with multiple environmental sensors, multiple output elements that provide notifications to users, and multiple notification event types may incorporate embodiments of the present application.
  • FIG. 1B is a block diagram illustrating components of a mobile computing device such as the computing device shown in FIG. 1A used in one embodiment. That is, the mobile computing device 100 may incorporate the system 102 to implement certain embodiments.
  • the system 102 can be used to implement a "smart phone" that can run one or more applications similar to desktop or notebook computer applications, such as presentation applications, browsers, email, scheduling, instant messaging , And media player applications.
  • the system 102 is integrated as a computing device, such as an integrated personal digital assistant (PDA) and wireless phone.
  • PDA personal digital assistant
  • One or more applications 166 may be loaded into memory 162 and run on or in association with operating system 164. Examples of application programs include telephone dialing programs, e-mail programs, PIM (Personal Information Management) programs, word processing programs, spreadsheet programs, Internet browser programs, message communication programs, and so on.
  • the system 102 also includes non-volatile storage 168 within the memory 162. The non-volatile storage 168 may be used to store persistent information that will not be lost when the system 102 is powered off.
  • the application 166 may use the information and store the information in the non-volatile storage 168, such as email or other messages used by the email application.
  • Synchronization applications may also reside on the system 102 and be programmed to interact with corresponding synchronization applications resident on the host computer to maintain the information stored in the non-volatile storage 168 and the host Corresponding information on the computer is synchronized. As should be understood, other applications may be loaded into the memory 162 and run on the device 100, including the auto insurance claims application 26.
  • the system 102 has a power source 170 that can be implemented as one or more batteries.
  • the power supply 170 may also include an external power source, such as a supplementary battery or an AC adapter to recharge the battery or a powered docking station.
  • the system 102 may also include a radio 172 that performs the function of transmitting and receiving radio frequency communications.
  • the radio 172 facilitates the wireless connection between the system 102 and the "outside world" through a communication operator or service provider.
  • the transmission to and from the radio 172 is performed under the control of the operating system 164. In other words, the communication received by the radio 172 can be propagated to the application 166 through the operating system 164 and vice versa.
  • Radio 172 allows system 102 to communicate with other computing devices, for example, over a network.
  • Radio 172 is an example of a communication medium.
  • the communication medium is embodied by computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and includes any information delivery medium.
  • modulated data signal refers to a signal that causes one or more of its characteristics to be set or changed in such a manner as to encode information in the signal.
  • communication media includes wired media such as a wired network or direct-wire connection, and wireless media such as acoustic, RF, infrared, and other wireless media.
  • the term computer-readable media as used herein includes both storage media and communication media.
  • This embodiment of the system 102 is shown with two types of notification output devices: an LED 120 that can be used to provide visual notifications, and an audio interface 174 that can be used to provide audio notifications from the speaker 125. These devices may be directly coupled to the power supply 170 so that when the system is activated, even if the processor 160 and other components may be turned off to save battery power, they remain powered on for a duration indicated by the notification mechanism.
  • the LED 120 can be programmed to remain powered on indefinitely until the user takes action to indicate the power-on status of the device.
  • the audio interface 174 is used to provide auditory signals to the user and receive auditory signals from the user.
  • the audio interface 174 may also be coupled to a microphone to receive auditory input, such as to facilitate telephone conversations.
  • the microphone may also serve as an audio sensor to facilitate the control of notifications, as will be described below.
  • the system 102 may further include a video interface 176 that allows the operation of the onboard camera 130 to record still images, video streams, and the like.
  • the mobile computing device implementation system 102 may have additional features or functions.
  • the device may also include additional data storage devices (removable and/or non-removable), such as magnetic disks, optical disks, or magnetic tape.
  • additional storage is shown by storage 168 in FIG. 1B.
  • Computer storage media may include volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules, or other data.
  • the data/information generated or captured by the device 100 and stored via the system 102 may be stored locally on the device 100 as described above, or the data may be stored in a separate device that can be connected to the device 100 by the device via radio 172 or through the device 100
  • the separate computing device is like a server computer in a distributed computing network such as the Internet.
  • data/information can be accessed via device 100, via radio 172, or via a distributed computing network.
  • these data/information can be easily transferred between computing devices for storage and use according to known data/information transmission and storage means, including email and collaborative data/information sharing systems.
  • FIG. 2 shows a networked environment in which embodiments of the present application can be implemented.
  • the user holds a tablet computing device 204 or a mobile computing device 206, and each of the tablet computing device 204 or the mobile computing device 206 includes the auto insurance claims application client 202 described herein.
  • the user can operate the auto insurance claim application client 202 and communicate with the auto insurance claim application server 212 through a network 208 including but not limited to the Internet.
  • the auto insurance claim application server 212 is implemented in the server 210 and includes a fixed loss model 214 for providing users with auto insurance claim services, specifically for image analysis, auto loss decision making, and calculation of claim amounts.
  • the auto insurance claim application server 212 also includes a storage 216 for storing various operation data or model data.
  • the auto insurance claim application server 212 also interfaces with N insurance companies 218 to receive various insurance data from various insurance companies (including but not limited to spare parts data, vehicle model data, compensation rate data, insurance record data, maintenance data, etc.), and Feedback claims to insurance companies.
  • various insurance companies including but not limited to spare parts data, vehicle model data, compensation rate data, insurance record data, maintenance data, etc.
  • Feedback claims to insurance companies.
  • FIG. 3 shows a block diagram of an auto insurance claim application 300 according to an embodiment of the present application.
  • the auto insurance claim application 300 is used to quickly provide a fixed loss solution for a user's car accident. This function is implemented by the combination of the auto insurance claim application client 302 and the auto insurance claim application server 310.
  • the auto insurance claim application client 302 is installed as an application (APP) on the mobile computing device or tablet computing device owned by the user, and the auto insurance claim application client 302 may also be installed on the user's mobile computing device or tablet computing device Of existing applications.
  • APP application
  • the auto insurance claim application client 302 includes a UI presentation module 304, an image acquisition module 306, and a communication module 308.
  • the UI presentation module 304 presents a graphical user interface to the user through the display of the user's mobile computing device or tablet computing device to guide the user to take a car damage image and upload the car damage image, obtain fixed loss information, and confirm claim information.
  • the image acquisition module 306 is used to acquire vehicle damage image information. Specifically, the image acquisition module 306 can guide the user to capture the vehicle damage image through the mobile computing device or tablet computing device installed with the auto insurance claim application client 302 through the UI presentation module , Including the full car image including license plate information, the distant view image of the damaged area, the close-up image of the damaged area, and the detailed image of the damaged area. If the user's vehicle has more than one damage, the user can be guided to take a long-distance image of the damaged area, a close-up image of the damaged area, and a detailed image of the damaged area for each damage to completely record all the damage information.
  • the image acquisition module 306 can also determine whether the image captured by the user meets the requirements for automatic damage recognition, that is, determine whether the image captured by the user is available for the auto insurance claim application to automatically identify the car damage information. If it is determined that the image taken by the user does not meet the requirements of automatic damage recognition, the user is guided to retake the corresponding image to meet the requirements of automatic damage recognition.
  • the image acquisition module 306 can also guide the user to take a video of the car damage, so that it can obtain richer car damage information for automatic loss determination.
  • the communication module 308 enables the auto insurance claims application client 302 to communicate with the auto insurance claims application server 310 to transmit the image acquired by the image acquisition module 306 to the auto insurance claims application server 310 for further processing.
  • the auto insurance claim application server terminal 310 resides at the server of the auto insurance claim service provider.
  • the auto insurance claim service may be provided by the auto insurance claim service provider.
  • the auto insurance claim application server 310 includes a communication module 312, an image recognition module 314, a fixed loss module 316, a credit review module 318, a payment module 320, and a fixed loss model 322.
  • the communication module 312 is used to receive vehicle damage image data from the vehicle insurance claim application client 302 and transfer these data to the image recognition module 314.
  • the image recognition module 314 is used to perform image recognition on the received vehicle damage image. Specifically, first filter the received car damage images, that is, filter out the images that do not meet the requirements based on whether the automatic damage recognition requirements are met, and notify the car insurance claims application client 302 through the communication module 312 to guide the user to re-take The corresponding image. After receiving all necessary vehicle damage images, the image recognition module 314 performs noise removal, component recognition, damage detection, cause determination, and degree determination on these vehicle damage images. These functions will be described in more detail below, and these functions are based on the fixed loss model 322 included in the auto insurance claims application server 310.
  • the training and learning of the fixed-loss model 322 is a major challenge.
  • deep learning and computer vision technology have made great progress, and some simple tasks (such as ImageNet classification) have even achieved higher accuracy than humans.
  • ImageNet classification some simple tasks
  • the algorithm The tough road is still difficult, and there are still few effective technical solutions in the field of auto insurance.
  • the algorithm needs to deal with a variety of complex lighting conditions, stains, water droplets, vehicle structure and other interference factors, and learn the key information that is effective for fixed loss from massive data.
  • the tens of millions of disorderly auto insurance historical damage pictures from various insurance companies are structured, data sorted, cleaned and necessary annotations.
  • the number of photos in this huge image database And the complexity of the label is an order of magnitude higher than the existing ImageNet.
  • the sorted, cleaned and annotated historical pictures of auto insurance fixed loss are provided to the fixed loss model 322 for learning and training on a software and hardware integrated heterogeneous machine learning platform based on ASIC, FPGA, GPU and other chip technologies.
  • the fixed-loss model 322 utilizes a large amount of existing historical fixed-loss data to perform model iterative learning for different vehicle models, colors, and lighting conditions, and finally can output more accurate component recognition results and target multiple degrees of scratches and deformations , The conclusion of the fixed damage of the components such as cracking and shedding.
  • the fixed loss model 322 uses a deep neural network to detect damaged parts of the vehicle and their regions in the image.
  • it can be based on convolutional neural networks (Convolutional Neural Networks, CNN) and regional proposal networks (Region Proposal Networks, RPN), combined with pooling layers, fully connected layers, etc.
  • the fixed-loss model 322 is constructed in advance, and then the fixed-loss model 322 is trained through a large number of sorted, cleaned, and annotated historical pictures of auto insurance fixed losses from various insurance companies to generate a deep neural network.
  • a fully connected layer Fully-Connected Layer, FC
  • pooling layer data normalization layer, etc.
  • Softmax probability output layer
  • Convolutional neural networks generally refer to convolutional layers as the main structure combined with other neural networks, such as activation layers, which are mainly used for image recognition.
  • the deep neural network described in this embodiment may include a convolutional layer and other important layers (such as a pooling layer, a data normalization layer, an activation layer, etc.), combined with a regional proposal network (RPN) to form and generate.
  • Convolutional neural networks usually combine two-dimensional discrete convolution operations in image processing with artificial neural networks. This convolution operation can be used to automatically extract features.
  • the regional suggestion network (RPN) can take the features extracted from an image (any size) as input (two-dimensional features extracted using a convolutional neural network), and output a set of rectangular target suggestion boxes, each of which has an object score.
  • the convolutional neural network (CNN) used may be referred to as a convolutional layer (CNN), and the region recommendation network (RPN) may be referred to as a region recommendation layer (RPN).
  • the fixed-loss model 322 can also be combined with a modified network model based on the convolutional neural network improved or the region recommendation network improved, and the generated deep convolutional nerve is constructed after training with sample data Network, namely the fixed loss model 322.
  • the models and algorithms used in the above embodiments may select the same model or algorithm.
  • various models and variants based on the convolutional neural network and the region suggestion network such as Faster R-CNN, Y0L0, Mask-FCN, etc.
  • the CNN can use any CNN model, such as ResNet, Inception, VGG, etc. and its variants.
  • the convolutional network part of the neural network can use mature network structures that have achieved good results in object recognition, such as Inception, ResNet and other networks.
  • the input is a picture
  • the output is a plurality of image areas containing damage parts and the corresponding damage classification (the damage classification is used to determine the type of damage) and confidence (confidence is a parameter that indicates the degree of authenticity of the damage type ).
  • Faster R-CNN, Y0L0, Mask-FCN, etc. all belong to the deep neural network including convolutional layers that can be used in this embodiment.
  • the fixed loss model 322 described herein can detect the damaged part, the type and degree of damage, and the location area of the damaged part in the part image in combination with the region suggestion layer and the CNN layer.
  • the image recognition module 314 then transfers the recognition result to the fixed loss module 316, which in turn transmits the fixed loss result or the detailed information of the fixed loss to the corresponding insurance company, that is, the insurance company where the user purchased their car insurance , So that the insurance company can search for the OE code of the damaged component in the respective database based on the vehicle identification code (VIN code) (based on the license plate number), and then find the local repair and replacement price based on the OE code, and Combined with the fixed loss details from the image recognition module 314, a corresponding maintenance list is generated. The insurance company then sends the corresponding repair list back to the loss determination module 316.
  • the quotation of each insurance company may be different, and the final price may also be different.
  • the loss determination module 316 then transmits the repair list to the auto insurance claim application client 302 through the communication module 312, and the auto insurance claim application client 302 presents the repair list to the user through the UI presentation module 304.
  • FIG. 4 shows an example of a graphical user interface 400 presented by the UI presentation module 304 to the user. It can be understood that the embodiments of the present application are not limited to this exemplary user interface 400.
  • the UI presentation module 304 displays the total amount of the local accident fixed loss and the repair list to the user. If the user approves the repair list, the user can activate the "accept" button 404 to continue to the next step, the credit review step .
  • re-shoot If the user does not approve the repair list (for example, if the damage is not recognized, or feels that more expensive repairs should be performed), he can choose to re-shoot, that is, the user can activate the "re-shoot" button 406 to return to the car damage image shooting step It is used to retake the image of car damage and re-determine the damage. Alternatively, users can abandon automatic claims and transfer to manual processing, just like the traditional auto insurance reporting and claims process.
  • the user can also activate the "Premium Premium Forecast" button 402 to the right of the total amount of the fixed loss to predict the next year's premium through the auto insurance claims application based on factors such as local insurance, the number of risks in the current year, the amount of claims, and personal credit.
  • Personal credit here can be given by a third-party credit service provider, and personal credit includes but is not limited to auto insurance points, which will be described in more detail below.
  • the user interface 400 shown in FIG. 4 is for illustrative purposes only and is exemplary and non-limiting.
  • the credit review module 318 reviews the user's personal credit to determine whether to perform automatic instant payment or transfer to manual processing.
  • the auto insurance claim application server 310 uses the auto insurance points to determine the user's personal credit points, but it should be understood that other credit systems or grades, such as sesame credits, can also be used.
  • the car insurance score can also be based on people (such as gender, age, occupation, identity), behavior (such as the number of violations, whether you often go on high speeds, credit history, consumption habits, driving habits) and use environment (such as road type, road congestion)
  • people such as gender, age, occupation, identity
  • behavior such as the number of violations, whether you often go on high speeds, credit history, consumption habits, driving habits
  • use environment such as road type, road congestion
  • the car insurance score is obtained through accurate portrait and risk analysis of the car owner, for example, a score ranging from 300 to 700. The higher the score, the lower the risk. Low-risk car owners generally have good driving habits and credit behavior.
  • the insurance company can query the user's auto insurance standard scores, or combine their own data to model the labels to obtain their own auto insurance special scores, so as to make fairer auto insurance pricing based on the auto insurance scores.
  • the vehicle insurance claims data accumulated by insurance companies and the user profile data accumulated by credit service providers (all desensitized) are merged at an improved model iteration speed. Process) to generate the user's car insurance points.
  • the credit review module 318 may notify the compensation module 320 to perform automatic instant compensation, that is, automatically transfer the total amount of the fixed loss to the user without manual review Designated bank account or online account.
  • a sufficiently high credit may refer to a credit or one of its manifestations reaching or exceeding a certain threshold or level, such as a car insurance score exceeding 600.
  • the insurance company will reduce the credit of the vehicle owner to prevent the vehicle owner from continuing to commit fraud.
  • insurance companies can also cooperate with third-party neutral platforms to transfer the relevant information of this claim to the third-party platform after desensitization processing in accordance with the legal requirements for the collection of credit records across insurance companies, and in multiple insurance companies Share between to prevent car owners from making fraud claims in multiple insurance companies.
  • the third-party platforms described here include but are not limited to China Credit Guarantee.
  • the credit review module 318 notifies the compensation module 320 to convert the user's current claim process to manual processing, just like the traditional auto insurance claim process.
  • 5A and 5B show another example of an application for auto insurance claims according to another embodiment of the present application.
  • the user instead of taking a car damage image and transmitting it to the server for technical processing and then returning to the client, the user can also be guided to take videos around the car to provide more rich vehicle image information, and Thanks to the substantial improvement in the processor performance and image processing capabilities of smartphones in recent years, the functions of the auto insurance claims application server 310 can be integrated into the auto insurance claims application client 302 so that the client 302 can automatically do damage to the car. Real-time analysis.
  • AR augmented reality
  • the auto insurance claim application can guide the user to take a higher-quality video image in real time when the user shoots a video, such as guiding the user closer to the injury site to make a more accurate determination of the car damage.
  • FIG. 6 shows a detailed functional block diagram of the image recognition module 314 according to an embodiment of the present application.
  • the image recognition module 314 first filters the received car damage image through the image filtering component 602, that is, filters out images that do not meet the requirements based on whether the automatic damage recognition requirements are met, and through the communication module 312 to notify the auto insurance claim application client 302 to guide the user to retake the corresponding image.
  • the image filtering component 602 in the image recognition module 314 may determine whether the panoramic image includes a license plate, whether the view is too far, and so on.
  • the image filtering component 602 may determine whether the car detail image completely includes the car damage details, whether the details are clearly visible, and so on. If the uploaded image does not meet the requirements for automatic damage recognition, the image filtering component 602 filters out the image and guides the user to retake the corresponding image through the UI presentation module 304.
  • the noise removal component 604 in the image recognition module 314 uses the fixed-loss model 322 to identify and exclude various types such as light reflections, shadows, reflections, stains, water droplets, vehicle structure, etc. Noise interference factors.
  • the fixed loss model 322 has been trained and learned through a large amount of historical fixed loss data (especially noisy image data), and thus can accurately identify and eliminate various interference factors. As shown in diagram 702 in FIG. 7A, the reflection noise removal based on the fixed loss model 322 is shown, and after being processed by the noise removal component 604, the reflection light on the left front fender of the vehicle is removed.
  • the component recognition component 606 recognizes various components of the vehicle, including the damaged components, through the fixed-loss model 322 based on the distant view image of the vehicle damage location.
  • component recognition refers to detecting the category and position of each component from the image of a car, such as identifying where the front cover, left front headlight, bumper, grille, etc. are in the picture.
  • Faster R-CNN faster regional convolutional neural network
  • other embodiments may also be combined with other Technology to complete part identification.
  • the damage detection component 608 detects the damage location and category through a fixed loss model 322 trained with a large number of samples.
  • the damaged part refers to the specific component damaged, and the damage category includes, for example, scratching, deformation, cracking, and falling off. If the shooting angle of the image containing the damaged part is not good, correct the angle. Specifically, since users generally do not undergo professional training, it is naturally impossible to ensure that every uploaded photo is taken with clarity and clarity, so sometimes it is necessary to correct the image, so as to better perform damage detection, cause judgment and damage degree determination. Correction techniques include but are not limited to projection-based methods, Hough transform-based, linear fitting-based, and Fourier transform. Diagram 704 in FIG. 7B shows an example of correction of the image containing the damaged part.
  • the cause judgment component 610 judges the cause of the damage through the fixed-loss model 322 trained by a large number of samples, such as single-vehicle scraping, sharps damage, double-vehicle scraping, double-vehicle collision, etc.
  • the cause judging component 610 can also identify the suspected non-injury damage, for example, based on the color of the damage being significantly different or the damage location not matching the collision location, etc.
  • the degree determination component 612 determines the degree of damage through a fixed loss model 322 trained by a large number of samples, such as slight scratches (no primer), severe scratches (primer exposed), slight deformation, severe deformation, and slight cracking , Severe cracking, scrapping, etc., in order to generate the corresponding maintenance list.
  • the fixed loss model 322 described here has been described in detail above, so it will not be repeated here.
  • the embodiments of the present application are not limited to the aforementioned degree of damage.
  • FIG. 8 shows a flowchart of a method 800 for auto insurance auto payment according to an embodiment of the present application.
  • the steps shown in FIG. 8 may be implemented by hardware (eg, processor, engine, memory, circuit), software (eg, operating system, applications, drivers, machine/processor executable instructions), or a combination thereof To execute.
  • hardware eg, processor, engine, memory, circuit
  • software eg, operating system, applications, drivers, machine/processor executable instructions
  • various embodiments may include more or fewer steps than shown.
  • the user is guided to take a vehicle image and acquire the vehicle image.
  • the UI presentation module in the auto insurance claim application client 302 can be used to guide the user to take a photo of the car damage image through the mobile computing device or tablet computing device installed with the auto insurance claim application client 302, including the whole car image including the license plate information, the car damage A long-distance image of a part, a close-up image of a car damaged part, and a detailed image of a car damage.
  • the user can also be guided to shoot the car damage video, so as to enable obtaining richer car damage information to facilitate automatic real-time loss determination.
  • image recognition is performed on the acquired vehicle image to identify vehicle damage.
  • Image recognition includes image filtering, noise removal, component recognition, damage detection, cause judgment, and degree judgment. These functional steps will be described in more detail in FIG. 9. In one embodiment of the present application, these image recognition functions are completed by a fixed-loss model trained with a large number of samples based on a deep neural network.
  • a repair list is generated based on the vehicle damage.
  • Vehicle damage includes damage location, damage category and damage degree. Paint repair can be used for mild scratching of parts, sheet metal repair can be used for deformation of parts, and parts can be directly replaced for severe damage such as cracking and shedding of parts.
  • the insurance company After transmitting the fixed loss result or detailed loss information to the corresponding insurance company, the insurance company searches for the OE code of the damaged part in the respective database based on the vehicle identification code (VIN code) (based on the license plate number), and then OE code can find the price of local parts repair and replacement, and generate the corresponding repair list.
  • VIN code vehicle identification code
  • the user is asked whether to accept the generated repair list. If the user accepts, the flow continues to block 810, otherwise the flow returns to block 802 to redirect the user to take a vehicle image.
  • the user's personal credit is reviewed.
  • Personal credit can be combined with claim data provided by insurance companies and personal credit data provided by third-party companies.
  • the user's personal credit may take the form of car insurance points, but it should be understood that other forms of credit evaluation may be used.
  • block 812 it is determined whether the user's credit is high enough to make a determination whether to perform automatic instant payment or transfer to manual processing. If it is determined that the user's credit is sufficiently high, the flow continues to block 814 to automatically pay the user. Otherwise, the process continues to block 816, and the claim is processed manually by the traditional claim process.
  • FIG. 9 shows a flowchart of a method 900 for image recognition according to an embodiment of the present application.
  • the steps shown in FIG. 9 may be implemented by hardware (eg, processor, engine, memory, circuit), software (eg, operating system, application, driver, machine/processor executable instruction) or a combination thereof To execute.
  • hardware eg, processor, engine, memory, circuit
  • software eg, operating system, application, driver, machine/processor executable instruction
  • various embodiments may include more or fewer steps than shown.
  • the received car damage image is filtered.
  • the unqualified image is filtered based on whether the automatic damage recognition requirement is satisfied, and if it is determined that the image does not meet the requirement, the image is filtered out and the user is guided to retake the corresponding image.
  • noise removal is performed on the filtered car damage image. Specifically, various noise disturbance factors such as light reflections, shadows, reflections, stains, water droplets, vehicle structure, etc. are identified and eliminated.
  • various components of the vehicle are identified, including damaged components.
  • the logo is based on the long-distance image of the damaged part of the car, and detects the type and location of various components from the image of a car, such as identifying where the front cover, left front headlight, bumper, grille, etc. are in the picture.
  • the damaged location and category in the car damage image are detected to determine the specific damaged component and the damaged category.
  • the angle is corrected to enable better damage detection, cause judgment, and damage degree judgment.
  • the cause of the damage is determined.
  • Causes of damage include single-car scratching, sharps damage, double-car scratching, double-car collision, etc.
  • it is also possible to identify the suspected non-injury damage for example, based on the significantly different color of the damage or the damage location does not match the collision location, etc.
  • the degree of damage is determined in order to generate a corresponding repair list.
  • the degree of damage includes slight scratching, severe scratching, slight deformation, severe deformation, slight cracking, severe cracking, scrapping, etc.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Finance (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Operations Research (AREA)
  • Multimedia (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Technology Law (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

一种用于车险自动赔付的方法。该方法包括：引导用户拍摄车辆图像并获取所述车辆图像（802）；对所获取的车辆图像进行图像识别以标识车辆损伤（804）；基于所述车辆损伤来生成维修清单（806）；询问用户是否接受（808），如果是，审查用户信用（810）；以及判断用户信用是否足够高（812），如果是，则自动赔付所述用户（814）。

Description

一种车险自动赔付方法和系统 技术领域
本公开涉及金融科技，尤其涉及一种用于车险自动赔付的方法和系统。
背景技术
据统计，当今中国每年约有4500万件私家车保险索赔案，其中60％左右为“纯外观损伤”，即2700万件。目前，每单处理成本约为150元，这为各个保险公司带来了承重的查勘成本。
而且，车主发生事故时，车主通常会给保险公司打电话报案，由保险公司安排查勘人员到现场进行查勘定损，或由车主对损伤处进行拍照后将照片发给保险公司，在经过现场或远程查勘后，确定需要维修更换的项目，再由车主去维修厂进行修理，拿到发票或维修清单之后，将这些材料交给保险公司，保险公司对所有相关单据进行审核后，将维修费用赔付给车主。整个过程比较冗长，车主需要关注的事项较多，拿到赔付款的速度也很慢。
另外，为给车主提供便利，部分保险公司提供了代收单据的服务，但也需要保险公司花费相当的人工成本，也无法从根本上缩短车主拿到赔付款的时间。此外，国内保险行业也会采用小额案件快速赔付的方案，对于金额低于一定阈值的案件，提前将费用赔付给车主，之后再审核各类材料，但这又会带来一定的欺诈风险，对部分案件可能造成不合理赔付。
现有的赔付流程如果希望能够做到过程严谨控制，减少不合理赔付比例，就会使得赔付流程较长，手续繁琐，车主体验不好；如果希望减少车主麻烦，简化流程，快速赔付，又会提高不合理赔付比例。
当前，金融科技(英语：Financial technology，也称为FinTech)的快速发展使得解决车险行业的高查勘成本、低效赔付流程和理赔欺诈问题成为可能。金融科技是人工智能(AI)、深度学习等技术在金融产业取得突破性发展的大背景下应运而生的产物。如何在车险场景这片热土上，最大程度挖掘AI技术算法以及数据的价值和潜力，成为保险行业内一众玩家努力开垦的新方向。
因此，保险公司或相关企业通过运用科技手段(尤其是AI技术和算法)来使得理 赔流程变得更有效率并且降低查勘成本和不合理赔付率是合乎需要的。
发明内容
提供本申请内容来以简化形式介绍，并将在以下具体实施方式部分中进一步描述一些概念。本申请内容并不旨在标识出所要求保护的主题的关键特征或必要特征，也不旨在用于帮助确定所要求保护的主题的范围。
对保险公司而言，为了根据事故现场做出损伤原因的初步判断并确认损伤的车部件及损伤程度，往往要雇佣大量的人力长期驻派多地进行重复的拍照作为定损记录的依据。本申请提供的技术方案依托深度学习图像算法逐步替代定损环节中的重复性人工作业流程，将大大降低车险定损环节中的人力以及时间成本。
具体而言，针对保险理赔领域中的高查勘成本、理赔流程繁琐和理赔欺诈这些问题，本申请提供了一种车险自动赔付方案以解决上述缺陷。具体地，本申请希望通过图像识别、人工智能、深度学习、互联网支付技术及用户信用体系，提供一种针对纯外观案件全自动赔付的方法，既能实现赔付款实时到账，避免车主和查勘员参与的流程或手续过多，又能控制赔付的精准程度以防止欺诈行为。
在本申请的一个实施例中，提供了一种用于车险自动赔付的方法，该方法包括：
获取车辆图像；
对所获取的车辆图像进行图像识别以标识车辆损伤；
基于车辆损伤来生成维修清单；
审查用户信用；以及
自动赔付用户。
在本申请的一个实施例中，用户信用是在用户接受维修清单的情况下审查的，并且该方法还包括在用户不接受该维修清单的情况下重新获取车辆图像。在本申请的一个实施例中，自动赔付是在用户信用足够高的情况下执行的并且该方法还包括在用户信用不够高的情况下将用户的本次理赔流程转为人工处理。
在本申请的一个实施例中，提供了一种用于车险自动赔付的系统，该系统包括：
用于获取车辆图像的装置；
用于对所获取的车辆图像进行图像识别以标识车辆损伤的装置；
用于基于车辆损伤来生成维修清单的装置；
用于审查用户信用的装置；以及
用于自动赔付用户的装置。
在本申请的一个实施例中，用户信用是在用户接受维修清单的情况下审查的，并且该系统还包括用于在用户不接受该维修清单的情况下重新获取车辆图像的装置。在本申请的一个实施例中，自动赔付是在用户信用足够高的情况下执行的，并且该系统还包括用于在用户信用不够高的情况下将用户的本次理赔流程转为人工处理的装置。
在本申请的一个实施例中，提供了一种存储用于车险自动赔付的指令的计算机可读存储介质，这些指令包括：
用于获取车辆图像的指令；
用于对所获取的车辆图像进行图像识别以标识车辆损伤的指令；
用于基于车辆损伤来生成维修清单的指令；
用于审查用户信用的指令；以及
用于自动赔付用户的指令。
在本申请的一个实施例中，用户信用是在用户接受维修清单的情况下审查的，并且该计算机可读存储介质还包括用于在用户不接受该维修清单的情况下重新获取车辆图像的指令。在本申请的一个实施例中，自动赔付是在用户信用足够高的情况下执行的，并且该计算机可读存储介质还包括用于在用户信用不够高的情况下将用户的本次理赔流程转为人工处理的指令。
本公开的各方面一般包括如本文参照附图所描述并且通过附图所阐示的方法、装置、系统、计算机程序产品。
在结合附图研读了下文对本申请的具体示例性实施例的描述之后，本申请的其他方面、特征和实施例对于本领域普通技术人员将是明显的。尽管本申请的特征可能是针对某些实施例和附图来讨论的，但本申请的全部实施例可包括本文所讨论的有利特征中的一个或多个。换言之，尽管可能讨论了一个或多个实施例具有某些有利特征，但也可以根据本文讨论的本申请的各种实施例使用此类特征中的一个或多个特征。以类似方式，尽管示例性实施例在下文可能是作为设备、系统或方法实施例进行讨论的，但是应当领会，此类示例性实施例可以在各种设备、系统、和方法中实现。
附图说明
为了能详细理解本公开陈述的特征所用的方式，可参照各方面来对以上简要概述的内容进行更具体的描述，其中一些方面在附图中阐示。然而应该注意，附图仅阐示了本公开的某些典型方面，故不应被认为限定其范围，因为本描述可允许有其他等同有效的方面。
图1A、图1B和图2示出了其中可实施本申请的各实施例的各种操作环境。
图3示出了根据本申请的一个实施例的车险理赔应用的一个示例的框图。
图4示出了根据本申请的一个实施例的由UI呈现模块呈现给用户的图形用户界面的示例。
图5A和图5B示出了根据本申请的另一实施例的车险理赔应用的另一示例。
图6示出了根据本申请的一个实施例的图像识别模块的详细功能框图。
图7A示出了根据本申请的一个实施例的图像识别中的反光去除的示例。
图7B示出了根据本申请的一个实施例的图像识别中的角度矫正的示例。
图8示出了根据本申请的一个实施例的用于车险自动赔付的方法的流程图。
图9示出了根据本申请的一个实施例的用于图像识别的方法的流程图。
具体实施方式
以下将参考形成本申请一部分并示出各具体示例性实施例的附图更详尽地描述各个实施例。然而，各实施例可以以许多不同的形式来实现，并且不应将其解释为限制此处所阐述的各实施例；相反地，提供这些实施例以使得本公开变得透彻和完整，并且将这些实施例的范围完全传达给本领域普通技术人员。各实施例可按照方法、系统或设备来实施。因此，这些实施例可采用硬件实现形式、全软件实现形式或者结合软件和硬件方面的实现形式。因此，以下具体实施方式并非是局限性的。
图1A、图1B、图2及相关联的描述提供了其中可实施本申请的各实施例的各种操作环境的讨论。然而，关于图1A、图1B、图2所示出和讨论的设备和系统是用于示例和说明的目的，而非对可被用于实施本文所述的本申请的各实施例的大量计算设备配置的限制。
图1A和1B示出可用来实施本申请的各实施例的合适的移动计算环境，例如移动电话、智能电话、输入板个人计算机、膝上型计算机等。参考图1A，示出了用于实现各实施例的示例移动计算设备100。在一基本配置中，移动计算设备100是具有输入元件和输出元件两者的手持式计算机。输入元件可包括允许用户将信息输入到移动计算设备100中的触摸屏显示器105和输入按钮110。移动计算设备100还可结合允许进一步的用户输入的可选的侧面输入元件115。可选的侧面输入元件115可以是旋转开关、按钮、或任何其他类型的手动输入元件。在替代实施例中，移动计算设备100可结合更多或更少的输入元件。例如，在某些实施例中，显示器105可以不是触摸屏。在又一替代实施例中，移动计算设备是便携式电话系统，如具有显示器105和输入按钮110的蜂窝电话。移动计算设备100还可包括可选的小键盘135。可选的小键盘135可以是物理小键盘或者在触摸屏显示器上生成的“软”小键盘。
移动计算设备100结合输出元件，如可显示图形用户界面(GUI)的显示器105。其他输出元件包括扬声器125和LED 120。另外，移动计算设备100可包含振动模块(未示出)，该振动模块使得移动计算设备100振动以将事件通知给用户。在又一实施例中，移动计算设备100可结合耳机插孔(未示出)，用于提供另一手段来提供输出信号。
尽管此处组合移动计算设备100来描述，但在替代实施例中，本申请还可组合任何数量的计算机系统使用，如在台式环境中、膝上型或笔记本计算机系统、多处理器系统、基于微处理器或可编程消费电子产品、网络PC、小型计算机、大型计算机等。本申请的实施例也可在分布式计算环境中实践，其中由分布式计算环境中通过通信网络链接的远程处理设备来执行任务；程序可位于本机和远程存储器存储设备中。总而言之，具有多个环境传感器、向用户提供通知的多个输出元件和多个通知事件类型的任何计算机系统可结合本申请的实施例。
图1B是示出在一个实施例中使用的诸如图1A中所示的计算设备之类的移动计算设备的组件的框图。即，移动计算设备100可结合系统102以实现某些实施例。例如，系统102可被用于实现可运行与台式或笔记本计算机的应用类似的一个或多个应用的“智能电话”，这些应用例如演示文稿应用、浏览器、电子邮件、日程安排、即时消息收发、以及媒体播放器应用。在某些实施例中，系统102被集成为计算设备，诸如集成的个人数字助理(PDA)和无线电话。
一个或多个应用166可被加载到存储器162中并在操作系统164上或与操作系统164相关联地运行。应用程序的示例包括电话拨号程序、电子邮件程序、PIM(个人信 息管理)程序、文字处理程序、电子表格程序、因特网浏览器程序、消息通信程序等等。系统102还包括存储器162内的非易失性存储168。非易失性存储168可被用于存储在系统102断电时不会丢失的持久信息。应用166可使用信息并将信息存储在非易失性存储168中，如电子邮件应用使用的电子邮件或其他消息等。同步应用(未示出)也可驻留在系统102上并被编程为与驻留在主机计算机上的对应同步应用进行交互，以保持存储在非易失性存储168中的信息与存储在主机计算机上的对应信息相同步。如应被理解的，其他应用可被加载到存储器162中且在设备100上运行，包括车险理赔应用26。
系统102具有可被实现为一个或多个电池的电源170。电源170还可包括外部功率源，如补充电池或对电池重新充电的AC适配器或加电对接托架。
系统102还可包括执行发射和接收无线电频率通信的功能的无线电172。无线电172通过通信运营商或服务供应商方便了系统102与“外部世界”之间的无线连接。来往无线电172的传输是在操作系统164的控制下进行的。换言之，无线电172接收的通信可通过操作系统164传播到应用166，反之亦然。
无线电172允许系统102例如通过网络与其他计算设备通信。无线电172是通信介质的一个示例。通信介质由诸如载波或其他传输机制等已调制数据信号中的计算机可读指令、数据结构、程序模块或其他数据来体现，并包括任何信息传递介质。术语“已调制数据信号”是指使得以在信号中编码信息的方式来设置或改变其一个或多个特性的信号。作为示例而非限制，通信介质包括诸如有线网络或直接线连接之类的有线介质，以及诸如声学、RF、红外及其他无线介质之类的无线介质。如此处所使用的术语计算机可读介质包括存储介质和通信介质两者。
系统102的该实施例是以两种类型的通知输出设备来示出的：可被用于提供视觉通知的LED 120，以及可被用于扬声器125提供音频通知的音频接口174。这些设备可直接耦合到电源170，使得当系统被激活时，即使为了节省电池功率而可能关闭处理器160和其他组件，它们也在一段由通知机制指示的持续时间保持通电。LED 120可被编程为无限地保持通电，直到用户采取行动指示该设备的通电状态。音频接口174用于向用户提供听觉信号并从用户接收听觉信号。例如，除被耦合到扬声器125以外，音频接口174还可被耦合到话筒以接收听觉输入，诸如便于电话对话。根据各本申请的各实施例，话筒也可充当音频传感器来便于对通知的控制，如下文将描述的。系统102可进一步包括允许板载相机130的操作来记录静止图像、视频流等的视频接口176。
移动计算设备实现系统102可具有附加特征或功能。例如，该设备还可包括附加 数据存储设备(可移动的/或不可移动的)，诸如磁盘、光盘或磁带。此类附加存储在图1B中由存储168示出。计算机存储介质可包括以用于存储诸如计算机可读指令、数据结构、程序模块、或其他数据等信息的任何方法或技术实现的易失性和非易失性、可移动和不可移动介质。
设备100生成或捕捉的且经系统102存储的数据/信息可如上所述本地存储在设备100上，或数据可被存储在可由设备通过无线电172或通过设备100和与设备100相关联的分开的计算设备之间的有线连接访问的任何数量的存储介质上，该分开的计算设备如例如因特网之类的分布式计算网络中的服务器计算机。如应理解的，此类数据/信息可经设备100、经无线电172或经分布式计算网络来被访问。类似地，这些数据/信息可根据已知的数据/信息传送和存储手段来容易地在计算设备之间传送以存储和使用，这些手段包括电子邮件和协作数据/信息共享系统。
图2示出了其中可实现本申请的各实施例的联网环境。用户持有平板计算设备204或者移动计算设备206，平板计算设备204或者移动计算设备206中的每一者都包括本文描述的车险理赔应用客户端202。用户可操作车险理赔应用客户端202并通过包括但不限于因特网的网络208来与车险理赔应用服务器端212通信。
车险理赔应用服务器端212在服务器210中实现并且包括用于为用户提供车险理赔服务，具体而言用于进行图像分析、车损决策、理赔金额计算等操作的定损模型214。车险理赔应用服务器端212还包括用于存储各种操作数据或模型数据的存储216。
车险理赔应用服务器端212还与N个保险公司218对接以便从各个保险公司接收各种保险数据(包括但不限于零配件数据、车型数据、赔付率数据、出险记录数据、维修数据等等)以及向保险公司反馈理赔决策。
图3示出了根据本申请的一个实施例的车险理赔应用300的框图。该车险理赔应用300用于快速地为用户的车损事故提供定损方案，该功能通过车险理赔应用客户端302和车险理赔应用服务器端310的组合来实现。
车险理赔应用客户端302作为一个应用(APP)被安装在用户拥有的移动计算设备或平板计算设备上，并且该车险理赔应用客户端302也可以是用户的移动计算设备或平板计算设备上已安装的现有应用的插件。
车险理赔应用客户端302包括UI呈现模块304、图像获取模块306、以及通信模块308。UI呈现模块304通过用户的移动计算设备或平板计算设备的显示器来向用户呈 现图形用户界面以引导用户拍摄车损图像并上传车损图像、获得定损信息以及确认理赔信息。
图像获取模块306用于获取车损图像信息，具体而言该图像获取模块306可通过UI呈现模块来引导用户通过安装有车险理赔应用客户端302的移动计算设备或平板计算设备来拍摄车损图像，包括包含车牌信息的全车图像、车损部位的远景图像、车损部位的近景图像、以及车损细节图像。如果用户的车辆具有不止一处损坏，则可引导用户针对每一处损坏分别拍摄车损部位的远景图像、车损部位的近景图像、以及车损细节图像，以完整地记录全部车损信息。
另外，图像获取模块306还能够确定用户拍摄的图像是否满足自动损伤识别的要求，即确定用户拍摄的图像是否可供车险理赔应用自动识别车损信息。如果确定用户拍摄的图像不满足自动损伤识别的要求，则引导用户重新拍摄相应图像以达到自动损伤识别的要求。
在本申请的一个实施例中，图像获取模块306还能够引导用户拍摄车损视频，以使得能够获取更丰富的车损信息以便于自动定损。
通信模块308使得车险理赔应用客户端302能够与车险理赔应用服务器端310通信，以便将通过图像获取模块306获取的图像传送到车险理赔应用服务器端310以供进行进一步处理。
车险理赔应用服务器端310驻留在车险自动理赔服务提供商的服务器处，在本申请的一个实施例中，该车险自动理赔服务可由车险自动理赔服务提供商提供。
车险理赔应用服务器端310包括通信模块312、图像识别模块314、定损模块316、信用审查模块318、赔付模块320、以及定损模型322。
通信模块312用于从车险理赔应用客户端302接收车损图像数据并将这些数据传递至图像识别模块314。
图像识别模块314用于对所接收到的车损图像进行图像识别。具体而言，首先对接收到的车损图像进行过滤，即基于是否满足自动损伤识别要求来过滤掉不符合要求的图像，并且通过通信模块312来通知车险理赔应用客户端302以引导用户重新拍摄相应图像。图像识别模块314在接收到所有必需的车损图像后对这些车损图像进行噪声去除、部件识别、损伤检测、原因判断、以及程度判定。这些功能将在下文中更详细地描述，并且这些功能都基于车险理赔应用服务器端310所包含的定损模型322。
然而，在使用定损模型322来进行图像识别之前，该定损模型322的训练和学习是一项重大挑战。近年来，深度学习以及计算机视觉技术获得了长足发展，在部分简单的任务上(比如ImageNet分类)甚至达到了比人更高的精确度，但是面对车险定损这样一个复杂的现实场景，算法的攻坚道路仍是困难重重，在车险定损领域依然鲜见有效的技术方案落地。在现实场景中，算法需要应对多种复杂光照条件、污渍、水滴、车型构造等多种干扰因素，并从海量数据中学习到对定损有效的关键信息。
为此，在本申请的一个实施例中，首先对来自各个保险公司的千万级杂乱无章的车险定损历史图片进行结构化规整、数据整理、清洗以及必要的标注，这个庞大图像数据库的照片数量以及标签的复杂程度对比现有的ImageNet都要高出一个数量级。随后，将经整理、清洗和标注的车险定损历史图片提供给定损模型322以供在基于ASIC、FPGA、GPU等芯片技术的软硬一体化异构机器学习平台上进行学习和训练。该定损模型322利用已有的沉淀的大量历史定损数据,针对不同的车型、颜色和光照条件进行模型迭代学习，最终能够输出较为精准的部件识别结果以及针对多种程度的刮擦、变形、部件的开裂和脱落等损伤的定损结论。
具体而言，定损模型322使用深度神经网络检测车辆受损部位及其在图像中的区域。在本申请的一个示例性且非限制性实施例中，可以基于卷积神经网络(Convolutional Neural Network，CNN)和区域建议网络(Region Proposal Network，RPN)，结合池化层、全连接层等来预先构建该定损模型322，然后通过来自各保险公司的大量经整理、清洗和标注的车险定损历史图片来训练该定损模型322以生成深度神经网络。在其它实施例中，还可结合全连接层(Fully-Connected Layer，FC)、池化层、数据归一化层等。在另一实施例中，如果需要对受损部位进行分类，还可以在定损模型322中加入概率输出层(Softmax)等。
卷积神经网络一般指以卷积层为主要结构并结合其他如激活层等组成的神经网络，主要用于图像识别。本实施例中所述的深度神经网络可以包括卷积层和其他重要的层(如池化层，数据归一化层，激活层等)，并结合区域建议网络(RPN)共同组建生成。卷积神经网络通常是将图像处理中的二维离散卷积运算和人工神经网络相结合。这种卷积运算可以用于自动提取特征。区域建议网络(RPN)可以将一个图像(任意大小)提取的特征作为输入(可以使用卷积神经网络提取的二维特征)，输出矩形目标建议框的集合，每个框有一个对象的得分。在另一实施例中，可以将所使用的卷积神经网络(CNN)称为卷积层(CNN)、将区域建议网络(RPN)称为区域建议层(RPN)。在本申请的 其它实施例中，定损模型322还可以结合基于所述卷积神经网络改进后的或区域建议网络改进后的变种网络模型，经过样本数据训练后构建所生成的深度卷积神经网络，即定损模型322。
上述实施例中使用的模型和算法可以选择同类模型或者算法。具体地，例如在定损模型322中，可以使用基于卷积神经网络和区域建议网络的多种模型和变种，如Faster R-CNN、Y0L0、Mask-FCN等。其中的卷积神经网络(CNN)可以用任意CNN模型，如ResNet、Inception、VGG等及其变种。通常神经网络中的卷积网络部分可以使用在物体识别取得较好效果的成熟网络结构，如Inception、ResNet等网络。在ResNet网络中，输入为一张图片，输出为多个含有损伤部位的图片区域以及对应的损伤分类(损伤分类用于确定损伤类型)和置信度(置信度为表示损伤类型真实性程度的参量)。Faster R-CNN、Y0L0、Mask-FCN等都是属于本实施例可以使用的包含卷积层的深度神经网络。本文描述的定损模型322能够结合区域建议层和CNN层检测出受损部位、损伤类型和程度、以及受损部位在所述部件图像中所处的位置区域。
回到图3，随后图像识别模块314将识别结果传递至定损模块316，该定损模块316进而将定损结果或定损明细信息传送到相应的保险公司，即用户购买其车险的保险公司，以便由该保险公司基于车辆识别码(VIN码)(基于车牌号获取)去各自的数据库中查找受损部件的OE码，然后根据OE码就能找到当地的部件维修和更换的价格，并结合来自图像识别模块314的定损明细信息来生成相应的维修清单。随后保险公司将相应的维修清单传送回定损模块316。每个保险公司的报价可能不一样，最后出的价格也可能不一样。
定损模块316然后将维修清单通过通信模块312传递至车险理赔应用客户端302，车险理赔应用客户端302通过UI呈现模块304向用户呈现该维修清单。
图4示出了由UI呈现模块304呈现给用户的图形用户界面400的示例。可以理解，本申请的各实施例不限于该示例性用户界面400。如图所示，UI呈现模块304向用户显示本地事故定损总金额以及维修清单，如果用户认可该维修清单，则该用户可激活“接受”按钮404以继续至下一步骤，即信用审查步骤。如果用户不认可该维修清单(例如有损伤未识别，或觉得应该进行花费更高的维修)，则可选择重新拍摄，即该用户可激活“重新拍摄”按钮406以返回到车损图像拍摄步骤以用于重新拍摄车损图像并重新定损。或者，用户可放弃自动理赔，转入人工处理，如同传统车险报案和理赔流程。
另外，用户还可激活定损总金额右侧的“来年保费预测”按钮402，以基于本地出 险、本年度出险次数、理赔金额、个人信用等因素来通过车险理赔应用预测下一年度的保费。此处的个人信用可以由第三方信用服务提供商来给出，且个人信用包括但不限于车险分，这将在下文中更详细地描述。如本领域技术人员可以理解的，图4所示的用户界面400仅仅是出于阐示的目的而是示例性且非限制性的。
在用户认可所呈现的维修清单，即在用户激活“接受”按钮404的情况下，信用审查模块318审查用户的个人信用以确定是进行自动即时赔付还是转人工处理。在本申请的一个实施例中，车险理赔应用服务器端310采用车险分来确定用户的个人信用分，但应理解，还可以采用其他信用体系或等级，诸如芝麻信用等。
传统上，要预测客户下一年的保险金额或赔付率，车险行业一般的做法是基于客户上一年度的出险次数、区域、车型、车价、使用性质等因素来计算，这些因素都属于车因素。而车险分还能够基于人(比如性别、年龄、职业、身份)、行为(诸如违章次数、是否经常上高速、信用历史、消费习惯、驾驶习惯)和使用环境(比如道路类型、道路拥挤状况)三大方面的因素来更精准地预测客户下一年的保险金额或赔付率，从而极大地丰富了车险中“与人相关的”数据维度。车险分通过对车主进行精准画像和风险分析得出例如300-700不等的分数，分数越高，风险越低。低风险的车主一般都有良好的驾驶习惯、信用行为。
保险公司在获得用户授权的情况下，可以查询用户的车险标准分，或是结合自身数据对标签进行加工建模，得到自己的车险专用分，从而依据车险分进行更为公平的车险定价。具体地，通过快速进行海量数据标签挖掘以及用于提升预测性能的机器学习算法来以提高的模型迭代速度融合保险公司累积的车险理赔数据以及信用服务提供商累积的用户画像数据(都经过脱敏处理)以产生用户的车险分。
回到对信用审查模块318的描述，当确定用户的信用足够高时，信用审查模块318可通知赔付模块320进行自动即时赔付，即在没有人工审核的情况下自动将定损总金额汇入用户指定的银行账户或线上账户。作为示例而非限制，信用足够高可以指信用或其某一表现形式达到或超过特定阈值或等级，诸如车险分超过600分。
在本申请的一个实施例中，如果完成自动赔付，但事后在审核材料阶段发现车主有欺诈行为，则保险公司将降低车主的信用，以防止车主持续进行欺诈。而且，保险公司也可与第三方中立平台合作，将此次索赔相关信息经过符合法律要求的脱敏处理后，传送到第三方平台以进行跨保险公司的信用记录征集，并在多个保险公司之间共享，以防止车主在多家保险公司进行欺诈索赔。此处描述的第三方平台包括但不限于中国保信。
如果确定用户信用分不够高，则信用审查模块318通知赔付模块320将用户本次理赔流程转为人工处理，如同传统的车险理赔流程。
图5A和图5B示出了根据本申请的另一实施例的车险理赔应用的另一示例。在本申请的另一实施例中，不同于拍摄车损图像并将其传送到服务器端做技术处理再返回到客户端，还可以引导用户围绕车辆拍摄视频以提供更丰富的车辆图像信息，并且得益于近年来智能手机的处理器性能和图像处理能力的大幅提高，可以将车险理赔应用服务器端310的功能整合到车险理赔应用客户端302以使得客户端302能够自动对车出现的损伤做出即时分析。具体而言，可以在用户拍摄视频时通过AR(增强现实)技术实时叠加损伤判定和维修信息，从而直接且即时地告诉用户损伤的程度，向用户自动推荐维修清单，从而能大幅提高拍摄引导、反馈的时效性，甚至做到实时定损。如图5A中的图示502所示，在用户拍摄视频时车险理赔应用能够实时地引导用户拍摄更高质量的视频图像，诸如引导用户更靠近损伤部位以便做出更精确的车损判定。另外，如图5B中的图示504所示，在用户拍摄视频时做出实时定损判断之际还能够结合损伤原因判断做出诸如“疑似非本案损伤”等更高阶的智能判定。而且，由于不将用户车辆图像信息上传至服务器端而是利用用户持有的移动计算设备或平板计算设备的本地CPU和GPU的能力，因此能一定程度上解决用户隐私、信息安全等方面的问题。
图6示出了根据本申请的一个实施例的图像识别模块314的详细功能框图。在本申请的一个实施例中，图像识别模块314首先通过图像过滤组件602对接收到的车损图像进行过滤，即基于是否满足自动损伤识别要求来过滤掉不符合要求的图像，并且通过通信模块312来通知车险理赔应用客户端302以引导用户重新拍摄相应图像。例如，当用户拍摄了车辆全景图像并上传后，图像识别模块314中的图像过滤组件602可确定该全景图像是否包括车牌、是否取景过远，等等。作为另一示例，当用户拍摄了车损细节图像并上传后，图像过滤组件602可确定该车损细节图像是否完整地包括车损细节、细节是否清晰可见，等等。如果上传的图像不符合自动损伤识别要求，则图像过滤组件602过滤掉该图像并通过UI呈现模块304引导用户重新拍摄相应图像。
当已经接收到自动损伤识别所必需的所有图像后，图像识别模块314中的噪声去除组件604使用定损模型322来标识并排除诸如光反射、阴影、倒影、污渍、水滴、车型构造等多种噪声干扰因素。定损模型322已经通过所沉淀的大量历史定损数据(尤其是带噪声的图像数据)进行训练和学习，并由此能够准确地标识出各种干扰因素并将其排除。如图7A中的图示702所示，示出了基于定损模型322的反光噪声去除，在经过 噪声去除组件604的处理后，车辆左前翼子板上的反光被去除。
然后，部件识别组件606基于车损部位的远景图像通过定损模型322来识别车辆的各个部件，包括受损部件。具体地，部件识别指的是从一辆车的图像中检测出各个部件的类别和位置，如识别图片中哪里是前机盖、左前大灯、保险杠、格栅等等。在本申请的一个实施例中，基于定损模型322结合Faster R-CNN(更快的区域性卷积神经网络)等网络技术来完成部件识别，但应理解在其他实施例中还可以结合其他技术来完成部件标识。
在标识出车辆的各个部件后，损伤检测组件608通过经大量样本训练的定损模型322来检测损伤部位和类别。损伤部位指的是受损的具体部件，损伤类别包括例如刮擦、变形、开裂、脱落等。如果包含损伤部位的图像的拍摄角度不佳，对角度进行矫正。具体地，由于用户一般不会经过专业训练，自然不可能保证每一张上传的照片都拍得端正明晰，因此有时需要对图像进行矫正，从而能够更好地进行损伤检测、原因判断和损伤程度判定。矫正技术包括但不限于基于投影的方法、基于Hough变换、基于线性拟合，以及傅里叶变换。图7B中的图示704示出了包含损伤部位的图像的矫正示例。
随后原因判断组件610通过经大量样本训练的定损模型322来判断损伤原因，诸如单车刮擦、锐器损伤、双车刮擦、双车碰撞等。另外，原因判断组件610还能识别出疑似非本案损伤，例如基于损伤颜色显著不同或者损伤部位与碰撞部位不符，等等。
最后，程度判定组件612通过经大量样本训练的定损模型322来判定受损程度，诸如轻微刮擦(不露底漆)、严重刮擦(露底漆)、轻微变形、严重变形、轻微开裂、严重开裂、报废，等等，以便生成相应的维修清单。此处描述的定损模型322已经在上文中详细描述，因此在此不再赘述。本申请的实施例不限于上述受损程度。
图8示出了根据本申请的一个实施例的用于车险自动赔付的方法800的流程图。在各实施例中，图8所示的步骤可通过硬件(例如，处理器、引擎、存储器、电路)、软件(例如，操作系统、应用、驱动器、机器/处理器可执行指令)或其组合来执行。如本领域普通技术人员将理解的，各实施例可以包括比示出的更多或更少的步骤。
在802，引导用户拍摄车辆图像并获取车辆图像。可通过车险理赔应用客户端302内的UI呈现模块来引导用户通过安装有车险理赔应用客户端302的移动计算设备或平板计算设备来拍摄车损图像，包括包含车牌信息的全车图像、车损部位的远景图像、车损部位的近景图像、以及车损细节图像。另外，还能够确定用户拍摄的图像是否满足自 动损伤识别的要求，并且在确定用户拍摄的图像不满足自动损伤识别要求的情况下引导用户重新拍摄相应图像以达到自动损伤识别要求。在本申请的另一实施例中，还能够引导用户拍摄车损视频，以使得能够获取更丰富的车损信息以便于自动实时定损。
在框804，对所获取的车辆图像进行图像识别以标识车辆损伤。图像识别包括图像过滤、噪声去除、部件识别、损伤检测、原因判断、以及程度判定，这些功能步骤将在图9中更详细地描述。在本申请的一个实施例中，这些图像识别功能通过基于深度神经网络的经大量样本训练的定损模型来完成。
在框806，基于车辆损伤来生成维修清单。车辆损伤包括损伤部位、损伤类别和损伤程度。针对部件轻度刮擦可采用喷漆维修，针对部件变形可采用板金维修，针对部件开裂、脱落等严重损伤可直接更换部件。在将定损结果或定损明细信息传送到相应的保险公司后，该保险公司基于车辆识别码(VIN码)(基于车牌号获取)去各自的数据库中查找受损部件的OE码，然后根据OE码就能找到当地的部件维修和更换的价格，并生成相应的维修清单。
在框808，询问用户是否接受所生成的维修清单。如果用户接受，则流程继续至框810，否则流程返回至框802以重新引导用户拍摄车辆图像。
在框810，审查用户的个人信用。个人信用可结合保险公司提供的理赔数据以及由第三方公司提供的个人信用数据。在本申请的一个实施例中，用户的个人信用可采取车险分的形式，但应理解可采取其它信用评估形式。
在框812，确定用户信用是否足够高以做出是进行自动即时赔付还是转人工处理的判断。如果确定用户信用足够高，则流程继续至框814，自动赔付用户。否则，流程继续至框816，将本次理赔专由按照传统理赔流程的人工处理。
图9示出了根据本申请的一个实施例的用于图像识别的方法900的流程图。在各实施例中，图9所示的步骤可通过硬件(例如，处理器、引擎、存储器、电路)、软件(例如，操作系统、应用、驱动器、机器/处理器可执行指令)或其组合来执行。如本领域普通技术人员将理解的，各实施例可以包括比示出的更多或更少的步骤。
在902，对接收到的车损图像进行过滤。换言之，基于是否满足自动损伤识别要求来过滤掉不符合要求的图像，并且在确定图像不符合要求的情况下过滤掉该图像并引导用户重新拍摄相应图像。
在904，对经过滤的车损图像进行噪声去除。具体地，诸如光反射、阴影、倒影、 污渍、水滴、车型构造等多种噪声干扰因素被标识并排除。
在906，标识车辆的各个部件，包括受损部件。该标识基于车损部位的远景图像，并且从一辆车的图像中检测出各个部件的类别和位置，如识别图片中哪里是前机盖、左前大灯、保险杠、格栅等等。
在908，检测车损图像中的损伤部位和类别以确定具体受损部件以及受损类别。在本申请的一个实施例中，如果包含损伤部位的图像的拍摄角度不佳，对角度进行矫正以使得能够更好地进行损伤检测、原因判断和损伤程度判定。
在910，确定损伤原因。损伤原因包括单车刮擦、锐器损伤、双车刮擦、双车碰撞等。另外，还能识别出疑似非本案损伤，例如基于损伤颜色显著不同或者损伤部位与碰撞部位不符，等等。
在912，判定受损程度以便生成相应的维修清单。受损程度包括轻微刮擦、严重刮擦、轻微变形、严重变形、轻微开裂、严重开裂、报废，等等。
以上参考根据本申请的实施例的方法、系统和计算机程序产品的框图和/或操作说明描述了本申请的实施例。框中所注明的各功能/动作可以按不同于任何流程图所示的次序出现。例如，取决于所涉及的功能/动作，连续示出的两个框实际上可以基本上同时执行，或者这些框有时可以按相反的次序来执行。
以上说明、示例和数据提供了对本申请的组成部分的制造和使用的全面描述。因为可以在不背离本申请的精神和范围的情况下做出本申请的许多实施例，所以本申请落在所附权利要求的范围内。

Claims (23)

  1. 一种用于车险自动赔付的方法，所述方法包括：
    获取车辆图像；
    对所获取的车辆图像进行图像识别以标识车辆损伤；
    基于所述车辆损伤来生成维修清单；
    审查用户信用；以及
    自动赔付用户。
  2. 如权利要求1所述的方法，其特征在于，所述车辆图像通过引导所述用户拍摄所述车辆图像来获取。
  3. 如权利要求1所述的方法，其特征在于，所述用户信用在所述用户接受所述维修清单的情况下审查，并且所述方法还包括在所述用户不接受所述维修清单的情况下重新获取车辆图像。
  4. 如权利要求1所述的方法，其特征在于，所述自动赔付是在所述用户信用足够高的情况下执行的，并且所述方法还包括在所述用户信用不够高的情况下将所述用户的本次理赔流程转为人工处理。
  5. 如权利要求1所述的方法，其特征在于，所述方法由包括车险理赔应用客户端和车险理赔应用服务器端的车险理赔应用来执行。
  6. 如权利要求1所述的方法，其特征在于，所述方法由包括只包括车险理赔应用客户端的车险理赔应用来执行。
  7. 如权利要求1所述的方法，其特征在于，所述车辆损伤的标识也能基于所述用户拍摄的车辆视频。
  8. 如权利要求1所述的方法，其特征在于，所述图像识别基于定损模型，所述定损模型通过深度神经网络来实现。
  9. 如权利要求8所述的方法，其特征在于，所述图像识别包括图像过滤、图像去噪、部件识别、损伤检测、原因判断、以及程度判定。
  10. 如权利要求1所述的方法，其特征在于，如果在完成所述自动赔付后在审核材料阶段发现所述用户有欺诈行为，则降低所述用户信用，并且将该信息传送到第三方平台以进行跨保险公司的信用记录征集并在多个保险公司之间共享。
  11. 如权利要求1所述的方法，其特征在于，所述用户信用基于来自保险公司的理赔数据以及来自第三方信用服务提供商的个人信用数据两者。
  12. 一种用于车险自动赔付的系统，所述系统包括：
    用于获取车辆图像的装置；
    用于对所获取的车辆图像进行图像识别以标识车辆损伤的装置；
    用于基于所述车辆损伤来生成维修清单的装置；
    用于审查用户信用的装置；以及
    用于自动赔付用户的装置。
  13. 如权利要求12所述的系统，其特征在于，所述车辆图像通过引导所述用户拍摄所述车辆图像来获取。
  14. 如权利要求12所述的系统，其特征在于，所述用户信用在所述用户接受所述维修清单的情况下审查，并且所述系统还包括用于在所述用户不接受所述维修清单的情况下重新获取车辆图像的装置。
  15. 如权利要求12所述的系统，其特征在于，所述自动赔付是在所述用户信用足够高的情况下执行的，并且所述系统还包括用于在所述用户信用不够高的情况下将所述用户的本次理赔流程转为人工处理的装置。
  16. 如权利要求12所述的系统，其特征在于，所述系统包括车险理赔应用客户端和车险理赔应用服务器端。
  17. 如权利要求12所述的系统，其特征在于，所述系统只包括车险理赔应用客户端。
  18. 如权利要求12所述的系统，其特征在于，所述车辆损伤的标识也能基于所述用户拍摄的车辆视频。
  19. 如权利要求12所述的系统，其特征在于，用于图像识别的装置基于定损模型，所述定损模型通过深度神经网络来实现。
  20. 如权利要求19所述的系统，其特征在于，所述用于图像识别的装置包括用于图像过滤的装置、用于图像去噪的装置、用于部件识别的装置、用于损伤检测的装置、用于原因判断的装置、以及用于程度判定的装置。
  21. 如权利要求12所述的系统，其特征在于，如果在完成所述自动赔付后在审核材料阶段发现所述用户有欺诈行为，则将降低所述用户信用，并且将该信息传送到第三方平台以进行跨保险公司的信用记录征集并在多个保险公司之间共享。
  22. 如权利要求12所述的系统，其特征在于，所述用户信用基于来自保险公司的理赔数据以及来自第三方信用服务提供商的个人信用数据两者。
  23. 一种包含指令的计算机可读存储介质，所述指令用于执行如权利要求1-11中的任一项所述的方法。
PCT/CN2019/123328 2018-12-25 2019-12-05 一种车险自动赔付方法和系统 WO2020134943A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811594050.1 2018-12-25
CN201811594050.1A CN110009508A (zh) 2018-12-25 2018-12-25 一种车险自动赔付方法和系统

Publications (1)

Publication Number Publication Date
WO2020134943A1 true WO2020134943A1 (zh) 2020-07-02

Family

ID=67165211

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/123328 WO2020134943A1 (zh) 2018-12-25 2019-12-05 一种车险自动赔付方法和系统

Country Status (3)

Country Link
CN (1) CN110009508A (zh)
TW (1) TWI710997B (zh)
WO (1) WO2020134943A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111784685A (zh) * 2020-07-17 2020-10-16 国网湖南省电力有限公司 一种基于云边协同检测的输电线路缺陷图像识别方法
CN112465018A (zh) * 2020-11-26 2021-03-09 深源恒际科技有限公司 一种基于深度学习的车辆视频定损系统的智能截图方法及系统
CN112733635A (zh) * 2020-12-28 2021-04-30 北京迈格威科技有限公司 对象识别方法、装置及电子设备
CN114529420A (zh) * 2022-03-08 2022-05-24 上海恒格信息科技有限公司 一种基于保险需求的智能保障专家系统

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110009508A (zh) * 2018-12-25 2019-07-12 阿里巴巴集团控股有限公司 一种车险自动赔付方法和系统
CN110443717A (zh) * 2019-07-16 2019-11-12 阿里巴巴集团控股有限公司 一种基于信用评估的保险理赔方法和系统
JP6714253B1 (ja) * 2019-08-30 2020-06-24 ゼアーウィンスリーサービス株式会社 車両用部材損傷修理判定システム、車両用部材損傷の保証申請の可否を判断する装置、車両用部材損傷の保証申請の可否を判断する装置の作動方法、及びプログラム。
CN111461905A (zh) * 2020-05-28 2020-07-28 深圳麦亚信科技股份有限公司 车险骗赔规避方法、装置、计算机设备及存储介质
TWI818181B (zh) * 2020-06-23 2023-10-11 新局數位科技有限公司 車體定損輔助系統及其實施方法
CN112036283A (zh) * 2020-08-25 2020-12-04 湖北经济学院 图像智能识别车辆定损方法
CN112037235B (zh) * 2020-08-27 2023-01-10 平安科技（深圳）有限公司 伤情图片自动化审核方法、装置、电子设备及存储介质
CN112446513A (zh) * 2020-12-15 2021-03-05 恒大恒驰新能源汽车研究院(上海)有限公司 车辆零部件检测方法及装置、计算机可读存储介质
CN112712498A (zh) * 2020-12-25 2021-04-27 北京百度网讯科技有限公司 移动终端执行的车辆定损方法、装置、移动终端、介质
CN115511124B (zh) * 2022-09-27 2023-04-18 上海网商电子商务有限公司 一种基于售后维修记录的客户分级方法
CN115410174B (zh) * 2022-11-01 2023-05-23 之江实验室 一种两阶段车险反欺诈图像采集质检方法、装置和系统
CN115631002B (zh) * 2022-12-08 2023-11-17 邦邦汽车销售服务(北京)有限公司 基于计算机视觉的车险智能定损方法及系统
CN115810134B (zh) * 2023-02-14 2023-07-21 之江实验室 一种车险反欺诈的图像采集质检方法、系统和装置
CN116935293B (zh) * 2023-09-15 2024-01-02 国任财产保险股份有限公司 一种远程视频查勘定损自动化处理方法及系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101925919A (zh) * 2007-11-28 2010-12-22 安信龙股份公司 自动化索赔处理系统
CN104268783A (zh) * 2014-05-30 2015-01-07 翱特信息系统（中国）有限公司 车辆定损估价的方法、装置和终端设备
CN107679996A (zh) * 2017-10-10 2018-02-09 民太安财产保险公估股份有限公司 车险理赔方法及系统
CN108009834A (zh) * 2017-12-27 2018-05-08 上海唯链信息科技有限公司 一种基于区块链技术的汽车保险信息系统
CN108280769A (zh) * 2018-02-01 2018-07-13 阿里巴巴集团控股有限公司 理赔业务的数据处理方法、装置、电子设备、服务器
CN108647712A (zh) * 2018-05-08 2018-10-12 阿里巴巴集团控股有限公司 车辆损伤识别的处理方法、处理设备、客户端及服务器
CN110009508A (zh) * 2018-12-25 2019-07-12 阿里巴巴集团控股有限公司 一种车险自动赔付方法和系统

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9886693B2 (en) * 2009-03-30 2018-02-06 Yuh-Shen Song Privacy protected anti identity theft and payment network
CN106780048A (zh) * 2016-11-28 2017-05-31 中国平安财产保险股份有限公司 一种智能车险的自助理赔方法、自助理赔装置及系统
CN107292528A (zh) * 2017-06-30 2017-10-24 阿里巴巴集团控股有限公司 车险风险预测方法、装置及服务器
CN108898504B (zh) * 2018-07-09 2021-12-07 北京精友世纪软件技术有限公司 一种移动查勘定损系统的智能训练及完善方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101925919A (zh) * 2007-11-28 2010-12-22 安信龙股份公司 自动化索赔处理系统
CN104268783A (zh) * 2014-05-30 2015-01-07 翱特信息系统（中国）有限公司 车辆定损估价的方法、装置和终端设备
CN107679996A (zh) * 2017-10-10 2018-02-09 民太安财产保险公估股份有限公司 车险理赔方法及系统
CN108009834A (zh) * 2017-12-27 2018-05-08 上海唯链信息科技有限公司 一种基于区块链技术的汽车保险信息系统
CN108280769A (zh) * 2018-02-01 2018-07-13 阿里巴巴集团控股有限公司 理赔业务的数据处理方法、装置、电子设备、服务器
CN108647712A (zh) * 2018-05-08 2018-10-12 阿里巴巴集团控股有限公司 车辆损伤识别的处理方法、处理设备、客户端及服务器
CN110009508A (zh) * 2018-12-25 2019-07-12 阿里巴巴集团控股有限公司 一种车险自动赔付方法和系统

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111784685A (zh) * 2020-07-17 2020-10-16 国网湖南省电力有限公司 一种基于云边协同检测的输电线路缺陷图像识别方法
CN111784685B (zh) * 2020-07-17 2023-08-18 国网湖南省电力有限公司 一种基于云边协同检测的输电线路缺陷图像识别方法
CN112465018A (zh) * 2020-11-26 2021-03-09 深源恒际科技有限公司 一种基于深度学习的车辆视频定损系统的智能截图方法及系统
CN112465018B (zh) * 2020-11-26 2024-02-02 深源恒际科技有限公司 一种基于深度学习的车辆视频定损系统的智能截图方法及系统
CN112733635A (zh) * 2020-12-28 2021-04-30 北京迈格威科技有限公司 对象识别方法、装置及电子设备
CN114529420A (zh) * 2022-03-08 2022-05-24 上海恒格信息科技有限公司 一种基于保险需求的智能保障专家系统

Also Published As

Publication number Publication date
TWI710997B (zh) 2020-11-21
TW202025060A (zh) 2020-07-01
CN110009508A (zh) 2019-07-12

Similar Documents

Publication Publication Date Title
WO2020134943A1 (zh) 一种车险自动赔付方法和系统
US11625791B1 (en) Feedback loop in mobile damage assessment and claims processing
US11580605B2 (en) Feedback loop in mobile damage assessment and claims processing
US11361385B2 (en) Application facilitated claims damage estimation
US11386503B2 (en) Processing insured items holistically with mobile damage assessment and claims processing
US10573012B1 (en) Three dimensional image scan for vehicle
US9818154B1 (en) System and method for electronic processing of vehicle transactions based on image detection of vehicle license plate
US11915321B2 (en) Configuration and transfer of image data using a mobile device
US10572944B1 (en) Claims damage estimation using enhanced display
US10157405B1 (en) Systems and methods for implementing machine vision and optical recognition
CN111652087B (zh) 验车方法、装置、电子设备和存储介质
US20170061459A1 (en) Augmented reality solution for price evaluation
US11380148B2 (en) 3D vehicle model data capturing and retrieving for vehicle inspection, service and maintenance
US20230230175A1 (en) Systems and methods for using tokenized icons to provide insurance policy quotes
WO2023178930A1 (zh) 图像识别方法、训练方法、装置、系统及存储介质
US20190102749A1 (en) Methods and systems of retail automotive check-in with a mobile device
CN113065908A (zh) 租赁方法、租赁装置、租赁平台及存储介质
WO2018102440A1 (en) System and method for electronic processing of vehicle transactions based on image detection of vehicle license plate
CN104102997A (zh) 一种信用卡流媒体识别技术并实现在线支付的方法
CN113327161A (zh) 一种用于信贷业务的智能风控决策系统
US20230154219A1 (en) Systems and methods for identifying a presence of a completed document
CN116911859A (zh) 一种手势支付方法、装置、设备和介质
CN113329121A (zh) 操作执行方法、操作执行装置、电子设备和可读存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19903994

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19903994

Country of ref document: EP

Kind code of ref document: A1